Beginner 📅 Last Updated: July 1, 2026 ⏱️ 15 min read

How to Install Open WebUI with Ollama

Who This Is For

Read this if: You have Ollama installed and working, and you want a ChatGPT-style web interface for your local models.

Skip if: Ollama is not installed yet. Start with How to Install Ollama.

What You Need

• Ollama installed and running (verify with ollama list)
• Docker installed (Docker Desktop on Windows/Mac, docker.io on Linux)
• At least one model pulled (e.g., ollama pull llama3.1:8b)
• 2GB free RAM for the container
• Time: 10–15 minutes

Step-by-Step Guide

Step 1: Verify Ollama Is Running

ollama list
# Should show your models

ollama ps
# Should show Ollama process running

If Ollama is not running, start it:

# Linux
sudo systemctl start ollama

# macOS / Windows
# Ollama app should be running in background

Step 2: Install Docker (If Not Already Installed)

Linux (Ubuntu):

sudo apt update
sudo apt install docker.io docker-compose
sudo systemctl enable docker
sudo systemctl start docker
sudo usermod -aG docker $USER
# Log out and back in for group change to take effect

Windows / macOS: Download Docker Desktop from docker.com and install.

Step 3: Run Open WebUI Container

docker run -d \
  -p 3000:8080 \
  --add-host=host.docker.internal:host-gateway \
  -v open-webui:/app/backend/data \
  --name open-webui \
  --restart always \
  ghcr.io/open-webui/open-webui:main

This downloads the Open WebUI image (~500MB) and starts it. The key part is --add-host=host.docker.internal:host-gateway — this allows the container to connect to Ollama running on your host machine.

Step 4: Open the Web Interface

Open your browser and go to:

http://localhost:3000

You will see the Open WebUI welcome screen. Create an admin account (this is stored locally — no data leaves your machine).

Step 5: Connect to Ollama

Open WebUI auto-detects Ollama in most cases. Verify in Settings → Connections:

• Ollama URL should be: http://host.docker.internal:11434
• Click the refresh icon — it should show a green "Connected" status

Step 6: Start Chatting

Select a model from the dropdown (your Ollama models appear automatically). Start chatting!

Common Mistakes & Errors

Error 1: "Connection Refused" — Ollama Not Detected

Cause: Docker container cannot reach Ollama on the host.

Fix:

# 1. Make sure Ollama is running
ollama ps

# 2. Check the connection URL in Open WebUI settings
#    Settings → Connections → Ollama URL
#    Must be: http://host.docker.internal:11434

# 3. If using Linux and host.docker.internal doesn't resolve:
#    Stop and recreate with network host mode:
docker stop open-webui && docker rm open-webui
docker run -d --network host -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:main
# Then Ollama URL becomes: http://127.0.0.1:11434

Error 2: "permission denied while trying to connect to Docker daemon"

Cause: Your user is not in the docker group.

Fix: sudo usermod -aG docker $USER then log out and back in.

Error 3: Port 3000 Already in Use

Cause: Something else is using port 3000.

Fix: Change the port mapping: -p 3001:8080 then access http://localhost:3001.

Error 4: Document Upload Not Working (RAG)

Cause: Missing embedding model or configuration.

Fix: Open WebUI handles RAG natively. Go to Settings → Documents and verify the embedding model is configured. It downloads automatically on first use.

Safe Defaults / Security Warning

Recommended Setup (Tiered)

What I Would Do

Run the default Docker command. Do not overcomplicate it. Open WebUI works perfectly with localhost binding and one admin account. Add the document chat feature — upload a PDF and ask questions about it. That is the killer feature that makes it feel like a real ChatGPT replacement.

Frequently Asked Questions

Should I install Open WebUI with Docker or pip?

Docker is recommended - handles dependencies and isolation cleanly: 'docker run -d -p 3000:8080 --add-host=host.docker.internal:host-gateway -v open-webui:/app/backend/data --name open-webui ghcr.io/open-webui/open-webui:main'. Pip works but requires manual dependency management.

How do I access Open WebUI from other devices?

It runs on port 3000 - access from another device at http://[your-ip]:3000. For Docker, map with '-p 3000:8080'. For external access, use a reverse proxy or Cloudflare Tunnel rather than opening router ports.

Does Open WebUI support multiple users?

Yes, built-in multi-user support with authentication. First account becomes admin, who can invite users and manage permissions. Each user gets separate chat history. Ideal for families, teams, or organizations.

How do I manage models in Open WebUI?

Open WebUI auto-detects Ollama models in the dropdown. Download new models from Settings > Models by entering the name (e.g., 'llama3.1'). Create custom presets with system prompts, temperature, and document attachments.

How do I update Open WebUI?

For Docker: 'docker pull ghcr.io/open-webui/open-webui:main' and recreate the container - data persists in the volume. For pip: 'pip install --upgrade open-webui'. Updates are frequent (often weekly).

Last Updated: July 1, 2026 — Verified against Open WebUI 0.4.x, Docker 27.x, Ollama 0.4.0.